Yuichi Inoue
Research Engineer at Sakana AI
Sagamihara, Japan
Summary
🎓
Top SchoolYuichi Inoue is a research engineer at Sakana AI with eight years of experience building agentic LLM systems focused on evaluation-driven development, inference-time scaling, and compute-efficient post-training. He led development of AB‑MCTS, an adaptive inference-time tree search framework (NeurIPS 2025 Spotlight), and co-created EvoVLM-JP v2, a Japanese vision-language model via evolutionary model merging. A Kaggle Competitions Grandmaster, he combines supervised fine-tuning and reinforcement learning to win top-10 results in large contests and recently helped his team secure a Gold Medal in the AI Mathematical Olympiad Progress Prize. Previously at Turing Inc., he worked on multimodal AI for autonomous driving, open-source training tooling, and real-time Jetson deployments, bridging research and production. Trained as a PhD scientist in pharmaceutical sciences at Kyoto University, he brings a rare blend of rigorous experimental science and competitive ML engineering to multimodal and reasoning-focused AI.
7 years of coding experience
5 years of employment as a software developer
京都大学薬学部薬科学科, 医薬創生情報科学専攻, 京都大学薬学部薬科学科, 医薬創生情報科学専攻 at 京都大学
Github Skills (18)
robotics10
racing9
perception8
rover8
autopilot7
sparkfun7
autonomous-vehicles7
iot7
gps4
python3
benchmark3
toolbox3
pytorch2
object-detection2
deep-learning2
Programming languages (3)
Github contributions (5)
turingmotors/heron
Sep 2023 - May 2024
Contributions:42 reviews, 22 PRs, 32 pushes in 7 months
Ino-Ichan/customized_mmdetection
Jul 2021 - Jul 2021
OpenMMLab Detection Toolbox and Benchmark
Contributions:17 commits, 16 pushes, 1 branch in 6 days
pytorchdeep-learningtoolboxobject-detectionbenchmark
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.